
KUALA LUMPUR - Singapore is developing a registry of artificial intelligence (AI) agents to support the safe use of AI by about 150,000 public officers across the government.
The registry will help authorities monitor the ownership and activities of AI agents, which are capable of making decisions and carrying out tasks with little human intervention.
The initiative forms part of a wider effort by the Government Technology Agency (GovTech) to strengthen oversight as public officers increasingly use AI tools for coding, report writing, research and scheduling tasks.
GovTech Chief Executive, Goh Wei Boon said the agency is also developing a platform known as AI Assistant Desk to provide government employees with a secure personal digital assistant.
“We want to have a layer of customisable rules, sanctioned AI tools and a registry to provide better visibility and security, so we can ensure that people use AI agents correctly,” said Goh.
AI agents can perform tasks in a way similar to human users by understanding instructions, carrying out online processes and making decisions through advanced language and reasoning capabilities.
The AI Assistant Desk is currently being tested by selected public officers ahead of a wider rollout later this year.
The project supports Singapore’s broader goal of expanding AI use across the public sector and improving service delivery.
The government has already been testing AI agents to help simplify business licensing applications and social assistance processes involving multiple agencies.
Goh said the platform's security features will remain in place even when external AI tools are integrated into the system.
These safeguards include restrictions that prevent AI agents from deleting files, sending emails to unauthorised recipients or generating inappropriate content.
GovTech has also introduced automated systems to ensure AI prompts and responses do not contain offensive language.
The agency said these measures are in line with its mission to help the public sector adopt technology in a secure and responsible manner.
Since its establishment in 2016, GovTech has grown from 1,800 employees to about 3,900 staff supporting more than 50 government agencies.
Besides maintaining government IT systems, the agency is leading efforts to integrate AI into the daily work of public officers.
“It is about bringing technology closer to the ground, enabling teams closest to real-world problems to develop more agile and impactful solutions.
“This evolution aims to transform agencies into stronger product owners by building deeper in-house capabilities,” said Goh.
GovTech currently deploys around 1,600 engineers across government agencies to provide technical support.
At the same time, AI-assisted coding and low-code tools are enabling more non-technical public officers to design and automate work processes.
The agency’s first hackathon in 2025 attracted about 600 public servants, with several projects progressing into real-world trials.
One of the projects is Markly, an AI-powered marking assistant currently being tested in 18 schools to help teachers assess handwritten examination papers more quickly and consistently.
Another project, LangBuddy, is an AI voice chatbot that helps students improve their Mandarin, Malay and Tamil language skills.
The tool is currently being tested by around 300 students from 10 secondary schools and junior colleges.
Singapore is also working towards its national target of producing 100,000 AI-fluent individuals by 2029 through specialised training programmes.
GovTech said more than half of the country's 150,000 public officers already use Pair, the government’s AI chatbot, to improve productivity and support research and writing tasks.
Cybersecurity remains a key part of Singapore’s AI strategy.
GovTech is developing AI-powered tools to conduct automated penetration testing on about 2,000 government systems that store citizen data and support public services.
The technology is designed to identify security weaknesses more efficiently than traditional manual testing methods.
“This process is very laborious. You can either do it manually yourself or engage a vendor. It can take a couple of months for each engagement, which is not very sustainable.
“If you ‘pen test’ yearly, you are only secure once a year, correct? With AI automation, we can test our systems continuously,” Goh added.
The push for stronger cybersecurity follows a cyber espionage attack targeting Singapore’s telecommunications sector in 2025.
Since January, GovTech has been using automated penetration testing tools and plans to expand their use across government systems.
The agency has also integrated AI-powered threat detection tools to help identify unusual network activity and emerging cyber threats.
“These developments are crucial to Singapore’s ongoing digital transformation to ensure better public service delivery and national security.
“With higher expectations and a faster-evolving threat environment, our work goes beyond mere maintenance,” said Goh.




